See related topic I just raised here: Existing numeric obs with no obs_reference_range
I would think we want to re-calculate, but those calculations need to take the data as of the obs datetime, not as of the current date time. But also looking for a clinician’s thoughts on this and the related issue I raised above, which could be solved by re-calculating reference ranges at the obs date time for all historically entered obs that never had a reference range calculated.